Contents lists available at ScienceDirect

Industrial Crops & Products

journal homepage: www.elsevier.com/locate/indcrop

Gene mining and identification of a flavone synthase II involved in flavonesbiosynthesis by transcriptomic analysis and targeted flavonoid profiling inChrysanthemum indicum L.

Yanfengyang Jianga,1, Xiaoyu Jia,1, Lixin Duanb,1, Peng Yea, Jinfen Yanga, Ruoting Zhana,Weiwen Chena,⁎, Dongming Maa,⁎

a Research Center of Chinese Herbal Resource Science and Engineering, Guangzhou University of Chinese Medicine, Key Laboratory of Chinese Medicinal Resource fromLingnan (Guangzhou University of Chinese Medicine), Ministry of Education, Joint Laboratory of National Engineering Research Center for the Pharmaceutics ofTraditional Chinese Medicines, Guangzhou, 510006, Chinab International Institute for Translational Chinese Medicine, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, China

A R T I C L E I N F O

Keywords:Chrysanthemum indicum L.Flavonoid profilingTranscriptomeFlavone synthase ⅡBiochemical function

A B S T R A C T

Chrysanthemum indicum L. is a type of herb that is widely used in China, Korea, and Japan. It has been used as aningredient in traditional medicines, tea, and functional food because of its various anti-inflammatory and anti-oxidant bioactivities. Such bioactivities have been associated with flavonoids such as apigenin, luteolin andlinarin in C. indicum. However, the biosynthesis pathway has not been investigated. In this study, using tran-scriptomic analysis and targeted metabolic profiling from five different tissues, we characterize the levels offlavonoids and mine the corresponding genes involved in flavonoid biosynthesis. Transcriptomic analysis re-vealed that 103 unigenes are involved in flavonoid-related biosynthesis pathways. Flavone synthase (FNS) is thekey enzyme responsible for flavone synthesis and provides precursors for acacetin and linarin biosynthesis. Oneputative FNS Ⅱ gene, with the highest Reads Per Kilobase per Million mapped reads (RPKM) in flower and flowerbud was cloned. Quantitative real-time polymerase chain reaction (RT-qPCR) revealed that CiFNSⅡ exhibited asimilar expression pattern to that in the transcriptome in terms of RPKM. In addition, a targeted metabolicprofiling of three flavanones (naringenin, eriodictyol, and liquiritigenin), three flavones (apigenin, luteolin, and7,4′-dihydroxyflavone), and two flavone derivatives (linarin and acacetin) was performed to characterize thedistribution of these flavonoids in different tissues of C. indicum. The recombinant FNSⅡ protein expressed inyeast was able to catalyze the conversion of three flavanones into the respective flavones. Based on the tran-scriptome analysis, metabolic profiling, and activity assays, a linarin biosynthesis pathway is proposed. Ourstudy provides insight into the potential application of molecular breeding and metabolic engineering for im-proving the quality of cultivated C. indicum.

1. Introduction

Chrysanthemum indicum L. (Asteraceae) is a native plant of Asia andNortheastern Europe. In China, the dried inflorescence of C. indicum hasbeen used as medicine for hundreds of years. Besides China, it also has along history of application in Korea and Japan as an ingredient fortraditional medicinal products, tea and functional foods because of itsmultiple bioactivities such as anti-inflammatory, anti-oxidant, anti-no-ciceptive, anti-bacterial, and anti-viral activities (Dong et al., 2017;Jeong et al., 2013; Kim et al., 2015; Luyen et al., 2015a; Nepali et al.,2018; Sun et al., 2016; Wu et al., 2013; Yang et al., 2017a, b; Yoo et al.,

2016; Zhang et al., 2015). In 2012, the Ministry of Health of the Peo-ple’s Republic of China listed C. indicum as an item in the list of func-tional food, which increased its potential market demand (Zhang et al.,2007).

Many bioactive components of C. indicum responsible for its phar-maceutical and bioactive activities have been reported. These includeflavonoids, essential oil, terpenoids and polysaccharides (Matsudaet al., 2002; Sun et al., 2015; Yoshikawa et al., 1999, 2000). Flavonoidsincluding acacetin, linarin (acacetin-7-O-rutinoside), apigenin, luteolinwere identified by HPLC analysis as the main active pharmacologicalcomponents of C. indicum responsible for its anti-inflammatory,

https://doi.org/10.1016/j.indcrop.2019.04.009Received 8 November 2018; Received in revised form 3 April 2019; Accepted 5 April 2019

⁎ Corresponding authors.E-mail addresses: chenww@gzucm.edu.cn (W. Chen), madm@gzucm.edu.cn (D. Ma).

1 These authors contributed equally to this paper.

Industrial Crops & Products 134 (2019) 244–256

0926-6690/ © 2019 Published by Elsevier B.V.

T

http://www.sciencedirect.com/science/journal/09266690https://www.elsevier.com/locate/indcrophttps://doi.org/10.1016/j.indcrop.2019.04.009https://doi.org/10.1016/j.indcrop.2019.04.009mailto:chenww@gzucm.edu.cnmailto:madm@gzucm.edu.cnhttps://doi.org/10.1016/j.indcrop.2019.04.009http://crossmark.crossref.org/dialog/?doi=10.1016/j.indcrop.2019.04.009&domain=pdf

immunoregulatory and analgesic effects (Kim et al., 2016, 2013; Leeet al., 2015; Luyen et al., 2015a; Wu et al., 2013). Although major focushas been placed on identifying the bioactivities of phytochemicals inthe previously-mentioned studies, genes responsible for the synthesis ofthe bioactive components in C. indicum remains unidentified.

Flavonoids, a group of valuable secondary metabolites, are ex-tensively distributed among plants and more than 10,000 flavonoidsdescribed to date (Mathesius, 2018). These phenolic compounds playimportant roles in plants, such as protecting tissues from UV damage,flower coloration and signals for interspecific interactions (Dixon andPasinetti, 2010; Mathesius, 2018; Xiao et al., 2016). All flavonoids aresynthetized from the precursor naringenin (Falcone Ferreyra et al.,2012). The synthesis of flavones, one of the most abundant flavonoidsubclass, is catalyzed by flavone synthase and is a rate-limiting step(Jiang et al., 2016). Based on their chemical structures, bioactive fla-vonoids in C. indicum, namely acacetin, linarin (acacetin-7-O-rutino-side), apigenin, luteolin, are classified as flavones or flavones deriva-tives.

FNS is responsible for the conversion of flavonones (e.g., nar-ingenin, eriodictyol, liquiritigenin etc.) to flavones (e.g., apigenin, lu-teolin, 7,4′-dihydroxyflavone, etc.) (Wu et al., 2016). FNS is encoded bytwo gene families, namely flavone synthase Ⅰ (FNSⅠ) and flavone syn-thase Ⅱ (FNSⅡ) (Britsch et al., 1981; Kochs and Grisebach, 1987; Stotzand Forkmann, 1981). FNSⅠ has been reported in rice, maize, Arabi-dopsis and in the Apiaceae genus (Ferreyra et al., 2015; Kim et al.,2008; Wang et al., 2018b), while the presence of FNSⅡ is extensive inplants. FNSⅡ enzymes are cytochrome P450 monooxygenases(CYP450s), FNSⅡ enzymes belonging to the CYP93B subfamily arepresent in dicot plants, while those belonging to the CYP93 G subfamilyare found in monocot plants (Akashi et al., 1998; Lam et al., 2014;Zhang et al., 2007). FNS enzymes have been characterized in Lonicerajaponica, Lonicera macranthoides, and Glycine max to convert flavonones(e.g., naringenin, eriodictyol, Liquiritigenin, etc.) to flavones (e.g.,apigenin, luteolin, 7,4′-dihydroxyflavone, etc.) (Fliegmann et al., 2010;Wu et al., 2016). However, little is known about FNS genes from C.indicum plants. This study aims to determine whether FNS genes encodeFNSI or FNSⅡ type enzymes, and to investigate the enzymatic reactionscatalyzed by FNS enzymes from C. indicum.

In this study, transcriptomic sequencing and analysis of five tissuetypes of C. indicum were performed to uncover candidate genes in-volved in flavonoid synthesis pathway. Of the five candidate flavonesynthases, one unigene with the higher RPKM in flower and flower budwas cloned and transcript levels of the unigene in different tissues weredetermined by RT-qPCR. The recombinant protein expressed in theWAT11 yeast strain was able to convert flavanones (naringenin, erio-dictyol, and liquiritigenin) to the respective flavones (apigenin, lu-teolin, and 7,4′-dihydroxyflavone) in in vitro activity assays. Combinedwith the fact that liquiritigenin and 7,4′-dihydroxyflavone were notdetected in any tissues of C. indicum, we provide evidence that CiFNSⅡmight be an active flavone synthase enzyme involved in the conversionof naringenin and eriodictyol to apigenin and luteolin, respectively.These results provide a good basis for pathway elucidation, molecularbreeding, and metabolic engineering for improving the quality of C.indicum.

2. Materials and methods

2.1. Materials

2.1.1. Chemical sourcesNaringenin (NAR), eriodictyol (ERI), liquiritigenin (LIQ), luteolin

(LUT), apigenin (API) were purchased from Sigma (Sigma, USA). 7,4′-dihydroxyflavone (DHF), linarin and acacetin were purchased fromShanghai yuanye Bio-Technology Co., Ltd, China.

2.1.2. Plant materialsChrysanthemum indicum L. was planted in the nursery field of

Guangzhou University of Chinese Medicine (Guangzhou City,Guangdong Province, China) in natural environment. The root, stem,leaf, flower and flower bud from healthy plants were collected from 2-year-old healthy plants in November and frozen at −80 ℃.

2.2. RNA extraction, reverse transcription, and sequencing

Total RNAs from the root, stem, leaf, flower, flower bud of C. in-dicum were isolated with Magen HiPure Plant RNA Mini Kit (Magen,China). We used a DNase I (GeneMark, China) treatment before cDNAsynthesis to remove gDNA. After extraction, the degradation and con-tamination of RNA were checked in 2% agarose gels. The concentrationof RNA was evaluated using an Agilent 2100 169 Bioanalyzer. RNAwith OD260/OD280 at 1.8–2.2 was used for further analysis. The RNAwas converted into first-strand cDNA using SuperScript Ⅲ (TransGen,China) and was ligated to Illumina sequencing adapters. Paired-endsequencing was done using an Illumina HiSeq™ 2000 system by GeneDenovo Biotechnology Co. (Guangzhou, China).

2.3. Sequence de novo assembly and gene annotation

The de novo assembly of RNA-Seq was performed using the Trinitysoftware (Haas et al., 2013). The expression level of the assembledcontigs was examined by Bowtie version 0.12.8, and normalizationusing RPKM values were calculated by a program containing the RSEMalgorithm (https://github.com/deweylab/RSEM), distributed by theTrinity group. The sequencing data were assessed by FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), per sequencequality scores, per sequence GC content and sequence duplication le-vels.

2.4. Differentially expressed gene analysis and screening of the unigenesinvolved in flavonoid biosynthesis

The analysis of gene expression levels in five different tissues of C.indicum was performed using the edgeR program (http://www.bioconductor.org/packages/release/bioc/html/edgeR.html). The iden-tified differentially expressed genes (DEGs) were analyzed based on aFDR (false discovery rate) of ≤0.05 and fold change of ≥2 to de-termine significant differences between the five tissues. The gene se-quences were subjected to KEGG pathway analysis by using the blastxprogram (E value threshold set at 10−5). After annotation by the KEGGdatabase, MEGA7 (Kumar et al., 2016) and iTOL (http://itol.embl.de/)were used for multiple sequence alignments and for preparing phylo-genetic analysis.

2.5. Cloning of CiFNSⅡ, sequence alignment and phylogenetic analysis

A candidate FNS gene, CiFNSⅡ, was chosen for cloning. The full-length open reading frame (ORF) for CiFNSⅡ was amplified usingPrimerSTAR max DNA polymerase (Takara, Japan) according to themanufacturer’s protocol. PCR conditions were as followed: 98℃, 1 min;98 ℃, 10 s, 55 ℃, 15 s, 72 ℃, 10 s, 30 cycles; 72 ℃, 5 min. The se-quences of gene-specific primers CiFNSⅡ-F and CiFNSⅡ-R were listed inTable S1. PCR product was cloned into pENTR/D vector (Invitrogen,USA), and the Trans1-T1 Phage resistant chemically Escherichia colicompetent cells (TransGen, China) were consequently transformed withthe recombinant plasmid pENTR/D-CiFNSⅡ. Subsequently, DNA se-quencing was performed. Motif analysis and sequence alignment wereperformed after confirming the sequence using DNAMAN version 7.0.The phylogenetic analysis was conducted by the maximum likelihoodmethod in MEGA 7 (Kumar et al., 2016).

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

245

https://github.com/deweylab/RSEMhttp://www.bioinformatics.babraham.ac.uk/projects/fastqc/http://www.bioinformatics.babraham.ac.uk/projects/fastqc/http://www.bioconductor.org/packages/release/bioc/html/edgeR.htmlhttp://www.bioconductor.org/packages/release/bioc/html/edgeR.htmlhttp://itol.embl.de/

2.6. RT-qPCR Experiments

RNA isolation and cDNA synthesis method were described pre-viously. RT-qPCR was carried out using EvaGreen 2× qPCR MasterMix(ABM, Canada) in Bio-Rad CFX96 Touch™ Real-Time PCR DetectionSystem (Bio-rad, USA) according to the manufacturer’s instructions.GAPDH (KC508619) and EF1α (KF305681) were used as referencegenes for normalization (Gu et al., 2011; Shen et al., 2010). The relativeexpression levels of CiFNSⅡ in different tissues were calculated usingthe 2−ΔΔCT method (Schmittgen and Livak, 2008), ΔΔCT = (Cttarget-CTref)target-(Cttarget-CTref)ref. Each reference gene was calculated sepa-rately, and the geometric mean (Gharbi et al., 2015) of the two geneswas used as the standardization result of the double reference genes.The primer sequences designed by Primer Premier 5.0 which used forRT-qPCR were shown in Table S2. The analysis of each tissue type wasperformed in biological replicates of three and technical replicates oftwo.

2.7. Construction of expression vector and Western blot analysis

Infusion cloning (Clontech, Japan) system was used to constructyeast expression vector in this study. The pENTR/D vector with theCiFNSⅡ construct was used as templates to amplify the CiFNSⅡ frag-ment. EcoRⅠ and SpeⅠ were chosen as digestion sites for the linearizationof expression vector of pESC-URA linearization. Then, we designed PCRamplification using internal primers pUF-CiFNSⅡ and pUR-CiFNSⅡ, asshown in Table S1, with 15 bp extensions that are complementary tothe ends of the linearized pESC-URA vector. In addition, we added sixhistidines between core sequence and vector extensions as his tag forwestern blot detection in the reverse primer. PCR procedures were asdescribed in Section 2.5. The PCR reaction mixture was incubated for15min at 50 ℃. Subsequently, Saccharomyces cerevisiae WAT 11 yeaststrain (Urban et al., 1997) was transformed with pESC-URA-CiFNSⅡusing a standard lithium acetate protocol (Gietz and Woods, 2002).

pESC-URA contains an URA3 marker for galactose-regulated ex-pression; this system can be induced by galactose and is repressed byglucose. After incubation for 24 h in SC-U culture medium containing2% glucose, transformant WAT 11 was isolated and resuspended in SC-U medium without sugar donor for depletion of glucose. Then the yeastcells were collected and resuspended in SC-U medium with 2% ga-lactose for further induction of the recombinant protein expression. Anequivalent of two OD600 cells from duplicate of the inductive culturewere harvested at seven time points (0 h, 4 h, 8 h, 16 h, 24 h, 32 h, 48 h)over a period of 48 h (Zheng et al., 2015). Harvested cells were re-suspended in 800 μL of dd H2O, 200 μL of lysis buffer (2M NaOH, 7.5%β-mercaptoethanol) was added, and the mix was incubated at 30 ℃ for5min. Cell suspensions were shaken for 15min with a 0.2-g glass bead(diameter 300 μM) and added to 150 μL of 55% trichloroacetic acidbefore being spun at 12,000 ×g; the total protein of the yeats cells wassubsequently extracted. His mouse monoclonal antibody (Transgene,China) was used to identify target proteins with a His epitope byWestern blotting (Mahmood and Yang, 2012).

2.8. In vitro CiFNSⅡ activity assays

The yeast strain WAT11 transformed with pESC-URA-CiFNSⅡ wasprecultured as previously described and harvested after growth for 8 h,which is the time with the highest cell concentration. The cells wereresuspended in pH 7.5 Tris-EDTA buffer (10mM Tris, 1 mM EDTA) anddisrupted using a high-pressure homogenizer with 20,000 psi at 4 ℃.The microsomes were precipitated with CaCl2 at a final concentrationof 18mM (Liu et al., 2018). The resulting mixture turned into emulsionafter 15-min of incubation on ice, which means that the microsomalfractions of the yeast cells were precipitated by CaCl2. Microsomes wereharvested through centrifugation at 4480 ×g and resuspended in100mM potassium phosphate buffer (pH=7.6). The enzyme assay for

CiFNSⅡ was conducted with a 1-mL final volume containing 100mMpotassium phosphate buffer (pH=7.6), 5mM NADPH, 2mM DTT,30 μg/mL substrate (naringenin, eridictyol, or liquiritigenin), and100 μg of total proteins. Protein concentration was determined byBradford reagent (Tiangen, China). After incubation for 30min at 30 °C,each reaction was terminated by two extractions with ethyl acetate(1 mL), vacuum dried, and redissolved in HPLC-grade methanol(500 μL) for HPLC-MS analysis.

2.9. The determination of flavonoid contents and HPLC-MS analysis

The root, stem, leaf, flower and flower bud of C. indicum wereground into powder in liquid nitrogen. In total, 500mg of powder wasprecisely weighed and transferred to a conical flask. 40mL of methanolwas added, and the mixture was extracted ultrasonically for 45min at55 ℃. Extracts were concentrated with a hypobaric drying method.1mL HPLC-grade methanol was used to re-dissolve the residues toobtain the extraction solution for HPLC-MS analysis. The identity offlavonoids detected in this study, including linarin, acacetin, luteolin,eriodictyol, liquiritigenin, apigenin, naringenin, and 7,4′-dihydroxy-flavone, were confirmed by HPLC-MS analysis using authentic stan-dards. The quantification of these flavonoids were performed based onby their corresponding standard curves. Each tissue type was analyzedin biological replicates of three and technical replicates of three in thisexperiment.

The HPLC-MS analysis was performed as follows: an Agilent 1290Series UHPLC system was coupled online with a hybrid quadrupoletime of flight (Q-TOF) mass spectrometer (6540B, Agilent Technologies,Inc., CA, USA). 2 μL of a filtered sample was applied to a reversed-phasecolumn (SB-C18 RRHD, 2.1× 100mm, 1.8 μm; Agilent) with an in-linefilter (1290 infinity in-line filter; Agilent). The system was operated inthe negative ion mode at the flow rate of 0.4mL/min using solvent A(water with 0.1% formic acid) and solvent B (acetonitrile with 0.1%formic acid). The gradient started from 28% B for 15min, followed by28% B to 100% B in 1min, held for 4min, and a post run time of 4minfor re-equilibration. Data were collected in the negative ESI mode se-parate runs on a Q-TOF (Agilent 6540B) operated in full scan modefrom 100 to 1100m/z. The capillary voltage was 3500 V with a scanrate of 3 scans per second; the nebulizer gas flow rate was 11 L/min;drying gas flow was 8 L/min; gas temperature was 300 ℃, and theskimmer voltage was 65 V.

3. Results

3.1. The de novo assembled transcriptome of C. indicum and annotation

To gain insight into flavone biosynthesis in C. indicum, next gen-eration sequencing was performed on RNA isolated root, stem, leaf,flower, and flower bud of C. indicum. After performing high-throughputsequencing using the Illumina HiSeq 2000 platform, we obtained97.69%–98.23% high quality reads from the raw reads in five differenttissues, and the GC percentage ranged from 42.85% to 43.27% (TableS3). After removing the redundant sequences, 89,206 unigenes weregenerated with an N50 of 1138 and an average length of 714 bp (TableS4). The length distribution and reads coverage are shown in Fig. S1. Itcan be seen that the unigenes with length ≥500 nt account over 50% ofthe total unigenes, which leads to high quality of sequence assembly.The RPKM values were applied to normalize and evaluate the expres-sion level of these assembled unigenes (Wagner et al., 2012). A total of54,171 unigenes were annotated with a cut-off E-value of 10−5 on fourdatabases including 53,751 (60.25%) on NCBI non-redundant (nr),37,908 (42.49%) on Swiss-Prot, 31,973 (35.84%) on KOG (EuKaryoticOrthologous Groups) and 20,604 (23.09%) on KEGG (Kyoto En-cyclopaedia of Genes and Genomes), respectively (Table S5). A venngraph of the four database annotation results is shown in Fig. S2a, in-dicated that 15,941 of them being annotated by the Nr, Swiss-prot, KOG

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

246

Liya下划线

and KEGG databases simultaneously, which is approximately 29.4% ofall the mapped unigenes. Based on the KOG annotation (Fig. S2b), mostunigenes (10,155) were annotated as ‘General function prediction only’.Additionally, 1761 unigenes were annotated as ‘Secondary metabolitesbiosynthesis, transport and catabolism classification’. In the GO (GeneOntology) similarity analyses shown in Fig. S3, 7473 unigenes wereannotated into 3 groups: Biological Process, Cellular Component andMolecular Function. In ‘Biological Process’, 4332 unigenes weregrouped as metabolic process, while 3615 unigenes were assigned tocellular process. Single-organism process ranked third, with 2863 uni-genes. In ‘Cellular component’, most of the unigenes were classified tocell and cell part. For ‘Molecular Function’ terms, it was clearly shownthat most of the unigenes were grouped as catalytic activity andbinding.

3.2. Identification of differentially expressed genes

To determine differentially expressed genes (DEGs) from five dif-ferent tissues of C. indicum, |log2FC|> 1and FDR≤ 0.05 was appliedas the screening thresholds. The number of DEGs in ‘Leaf vs Flowerbud’, ‘Leaf vs Flower’, ‘Leaf vs Stem’ and ‘Leaf vs Root’ were 39,335,36,589, 23,802, and 31488, respectively, with up-regulated DEGs of22,165, 18,406, 13,170 and 20,353, and down-regulated DEGs of17,170, 18,183 10,632 and 11,135, respectively (Fig. S4). Based on thisanalysis, in comparison with stems and roots, more differences in geneexpression exist in flower and flower bud when compared with leaf,indicating that the number of genes under active transcription in flowerbuds and flowers is more abundant than in other tissues.

3.3. Annotated genes involved in flavonoid biosynthesis

KEGG database provides us with a basic platform for systematicstudy of gene function acted in the metabolic networks of gene products(Kanehisa et al., 2012). Of the 11,290 unigenes enriched in KEGG da-tabase, 103 of these belonged to the flavonoid biosynthesis relatedpathways (Table 1). On the other hand, 1408 unigenes were annotatedby KEGG database as belonging to the secondary metabolite pathways(Fig. 1). The expression patterns of the unigenes related to flavonoidbiosynthesis (Table S6) in the five different tissues (Fig. 2a) were gen-erated by the heatmap package of the R software (Fig. 2b). In total, fiveunigenes were annotated and classified as part of the CYP93 family.Three unigenes were classified as part of the CYP93 A subfamily, andtwo unigenes—namely 0028351 and 0043683—were classified as partof the CYP93B subfamily (Table S7). Some members of the CYP93Bsubfamily have been previously reported as FNSⅡ and favanone 2-hy-droxylase (F2H) (Akashi et al., 1998, 1999; Martens and Forkmann,1999; Yan et al., 2014; Zhang et al., 2007). The expression level ofunigene0043683 in the flower and flower bud was much higher notonly than that of unigene0028351 but also than that of other tissuetypes (Table S7 and Fig. 2b). Furthermore, analysis using blastx(https://blast.ncbi.nlm.nih.gov/Blast.cgi) and ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder/) programs revealed that only theunigene0043683 contains a complete ORF. Therefore, unigene0043683was selected as a candidate gene for further cloning and biochemicalcharacterization.

3.4. Isolation of full-length ORF for CiFNSⅡ and sequence analysis

Based on sequence analysis, the unigene0043683 ORF encodes aprotein with 514 amino acids and a molecular mass of 58.6 kDa. For thecomparison of characteristic sequence signatures of FNSⅡs, we alignedthe amino acid sequences of unigene0043683 with 4 characterizedFNSⅡ enzymes: LjFNSⅡ-1.1 and LjFNSⅡ-2.1 from Lonicera japonica(AMQ91109.1,AMQ91111.1), LmFNSⅡ-1.1 from Lonicera macranthoides(AMQ91113.1), and CYP93B2 from Gerbera hybrida (AAD39549.1) byDNAMAN version 7 (Fig. 3a). It shares the highest identity (84.47%)with CYP93B2 at the amino acid level. Consequently, we named uni-gene0043683 as CiFNSⅡ (GenBank MK419957). We found that fourcytochrome P450-featured conserved motifs are present in the CiFNSⅡamino acid sequence using the 'Conserved Domains’ search at the NCBIwebsite (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi).The highly conserved proline-rich hinge region typically obeys theconsensus LPPPPXXXP motif, and is supposed to serve as a ‘hinge’ thatis required for proper orientation of P450 enzymes (Chapple, 1998).Furthermore, the oxygen binding pocket motif AATDTT is found to beessential for oxygen binding and activation (Hasemann et al., 1995).We also found an E-R-R triad residue that encodes a pocket lockingmotif that stabilizes the core structure of the meander region and the‘folding in’ of the heme (Chapple, 1998; Hasemann et al., 1995). Theheme-binding domain contains the signature sequence PFGXGRRXCPGis often highly conserved in all P450 polypeptides which gives theseenzymes the carbon monoxide-binding ability (Chapple, 1998). Aphylogenetic analysis was conducted by using the deduced amino acidsequence of CiFNSⅡ and 11 FNSII proteins from 9 different plant species(Table S8) to investigate the relationship between them (Fig. 3b). SomeFNSⅡ enzymes have been reported to catalyze the direct conversion offlavanones to flavones by introducing a double bond between the C-2and C-3 residues in flavanones, such as CYP93B2 from G. hybrida(Martens and Forkmann, 1999), CYP93B3 from Antirrhinum majus, andCYP93B4 from Torrentia (Akashi et al., 1999). However, F2H CYP450enzymes catalyze the hydroxylation of flavanones generating 2-hydro-xyflavanones; for example, in the Fabaceae family, it was demonstratedthat CYP93B1 from Glycyrrhiza echinata and CYP93B10/11 from Med-icago truncatula have F2H activity (Akashi et al., 1998; Zhang et al.,2007). CiFNSⅡ is clustered closely with CYP93B2 (Martens andForkmann, 1999) in the same clade as FNSⅡ (Ayabe and Akashi, 2006),CYP93B1 and CYP93B10 (Akashi et al.,3 1998; Zhang et al., 2007) areinvolved in 2-hydroxyflavanone biosynthesis and clustered togetherand separated from the FNSⅡs.

3.5. Expression analysis of the CiFNSⅡ in different tissues of C. indicum

To determine the expression pattern of CiFNSⅡ, RT-qPCR was per-formed on five different tissues of C. indicum. including root, stem, leaf,flower, and flower bud, with GAPDH and EF1α as reference genes.Higher transcript levels were detected in flower and flower bud than inother tissues (Fig. 4a–c). The expression pattern observed for CiFNSⅡ isalmost in agreement with the transcriptome data represented by RPKMvalue, which indicates that CiFNSⅡ might be the most active flavonesynthase in flower and flower bud tissues in C. indicum (Fig. 4d).

3.6. CiFNSⅡ heterologous protein expression and in vitro enzyme activityassays

To gain insight into the functions of CiFNSⅡ, cDNA was obtainedfrom C. indicum. A pair of primers were designed to amplify the ORF ofCiFNSⅡ (Table S1). To investigate the catalytic activity of the enzymeencoded by the isolated putative CiFNSⅡ gene, the ORF region wasinserted into the pESC-URA vector and transformed into the WAT11yeast strain. WAT11 cells expresses the Arabidopsis NADPH-cytochromeP450 reductase, which provides the reducing equivalents necessary forthe activity of plant CYP450 s, including FNSⅡ enzymes. Western blot

Table 1Distribution of unigenes involved in flavonoids related pathways.

Pathway All genes with pathwayannotation (11290)

Pathway ID

Flavonoid biosynthesis 87 (0.77%) ko00941Anthocyanin biosynthesis 10 (0.09%) ko00942Isoflavonoid biosynthesis 2 (0.02%) ko00943Flavone and flavonol

biosynthesis4 (0.04%) ko00944

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

247

https://blast.ncbi.nlm.nih.gov/Blast.cgihttps://www.ncbi.nlm.nih.gov/orffinder/https://www.ncbi.nlm.nih.gov/orffinder/http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=protein&doptcmdl=genbank&term=AMQ91109.1http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=AMQ91111http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=protein&doptcmdl=genbank&term=AMQ91113.1http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=protein&doptcmdl=genbank&term=AAD39549.1http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=nucleotide&doptcmdl=genbank&term=MK419957https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgihttp://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=CYP450

assay was performed to explore the protein expression level of CiFNSⅡin yeast. CiFNSⅡ-expressing yeast cells were cultured and sampled at sixdifferent time points during a 48-h period, after induction withGalactose. CiFNSⅡ was successfully expressed during the 48-h induc-tion period, with a maximum expression level at 8 h, as shown byWestern blot analysis (Fig. S5). Consequently, induction with galactosefor 8 h was selected for the collection of cells used for microsomespreparation.

Microsomes were prepared from harvested cells and tested in FNSⅡenzyme assays with three flavanone substrates, namely naringenin,eridictyol, and liquiritigenin. Reaction products were analyzed byHPLC-MS analysis. A new candidate product peak was detected fromthe products of the enzyme assays.

When narigenin was used as a substrate, the product showed amolecular ion of m/z 269 in the negative ion mode, which is close tonaringenin’s molecular ion of m/z 271 (Fig. 5a). This product exhibits

Fig. 1. Distribution of unigenes in different secondary metabolism categories based on KEGG classifications. Numbers follow pathway description in brackets indicatethe relative percentage.

Fig. 2. (a) Representative images of C. indicum tissues used for transcriptome and metabolic profiling analysis. (b) Expression pattern analysis of unigenes involved inflavonoid-related biosynthesis pathway.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

248

the same retention time and mass fragmentation pattern as an authenticstandard of apigenin. Therefore, we propose that the reaction product isapigenin, which has a m/z ratio of 2 less than that of naringenin. Si-milarly, eriodictyol (m/z 287) was converted luteolin (m/z 285) fol-lowing the enzyme reaction (Fig. 5b), while liquiritigenin (m/z 255)was converted to 7,4′-dihydroxyflavone (m/z 253) (Fig. 5c). The MS/MS fragmentation profiles were shown in Figure S6-S11. Therefore,CiFNSⅡ-expressing microsomes appear to metabolize eriodictyol, nar-ingenin, and liquiritigenin to luteolin, apigenin, and 7,4′-dihydroxy-flavone, respectively. In our test, moreover, 2-hydroxynaringenin (m/z287) and 2-hydroxyliquiritigenin (m/z 271) could be identified throughthe extracted ion peak (Fig. 5a,c), although the respective standardsamples were inaccessible to us. This supports the hypothesis that

CiFNSⅡ encodes a flavone synthase, catalyzing the conversion of fla-vanones to flavones, probably through 2-hydroxyflavanone inter-mediates. Thus, a first step seem to involve the hydroxylation of fla-vanones, followed by a second step of dehydration, as both 2-hydroxyflavanones and flavones could be identified without acidtreatment.

3.7. Profiling of Flavonoid Metabolites in different tissues of C. indicum

The biosynthesis of flavones (apigenin, luteolin and 7,4′-dihydrox-yflavone) from flavanones (naringenin, eridictyol, and liquiritigenin,respectively) is catalyzed by FNSⅡ enzymes (Fliegmann et al., 2010; Wuet al., 2016). Flavone derivatives (linarin and acacetin) are important

Fig. 3. (a) Multiple sequences alignment of the CiFNSⅡ protein and four other flavone synthases, Gerbera hybrida CYP93B2 (AAD39549.1), Lonicera japonica LjFNSII-1.1&LjFNSII-2.1 (AMQ91109.1,AMQ91111.1), Lonicera macranthoides LmFNSII-1.1 (AMQ91113.1). Alignment was carried out with gap penalty of 7 and K-tupe of 2by DNAMAN version 7. Light-blue, yellow, and pink coloration reflect 100%, 75% and 50% amino acid residues conservation. CYP450-specific conserved motifsincluding proline-rich membrane hinge LPPPPXXXP (Ⅰ), oxygen binding pocket motif AATDTT (Ⅱ), E-R-R triade motif (Ⅲ), and heme-binding domain PFGXGRR-XCPG (Ⅳ) are shown with black frame. (b) Maximum likelihood tree illustrating the phylogenetic relationship of CiFNSⅡ with flavone synthases from several plantspecies. Phylogenetic analysis was performed by MEGA 7.0 and test of phylogeny was carried out by bootstrap method with 500 replications. Substitution model wasbuilt by poisson correction data subset to use pairwise deletion. CiFNSⅡ is marked in red.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

249

pharmaceutical integredients. To explore the distribution of FNSⅡ-re-lated flavones and flavones derivatives such as linarin and acacetinaccumulated in C. indicum, we gathered methanolic extracts from root,stem, leaf, flower, and flower bud for HPLC-MS analysis. The results areshown in Fig. 6, Fig. S12 and Table 2. From eight targeted compounds(naringenin, apigenin, eridictyol, luteolin, liquiritigenin, 7,4′-dihy-droxyflavone, linarin, and acacetin), 6 flavonoids except liquiritigeninand 7,4′-dihydroxyflavone were found among the different tissues, andthe composition was very different from each other. The phytochemicalresults showed that some common constituents, such as linarin, lu-teolin, and apigenin, were present in all the tissues. Flower representthe organ with the most abundant flavonoid variety, i.e., six flavonoidsdetected in our experimentt. Notably, acacetin was not detected inflower bud. The total contents of flavonoids detected in flower andflower bud were higher than those of the other organs such as root,stem and leaf. It was obvious that luteolin and apigenin were the mostabundant metabolites among all the tissues, and they also exhibitedmuch higher concentration in flower and flower bud sections as com-pared with other tissues. The flower had the highest luteolin(27.635 μg/mg FW) and apigenin (2.353 μg/mg FW) levels. As luteolinand apigenin are produced by CiFNSⅡ, we hypothesize that the highercontents of luteolin and apigenin are possibly related to higher FNSⅡtranscript levels in C. indicum. This hypothesis is consistent with ourtranscriptomic data. It is interesting that the root but not flower has thehighest level of linarin (0.364 μg/mg FW), which are about 6-foldhigher than those in flower.

4. Discussion

In recent years, the integration of both gene transcription and me-tabolism analyses has been widely used to reveal the biosynthesispathways of the bioactive components, such as flavonoids, caffeoyl-quinic acids, and bornyl acetate in medicinal plants (Wang et al.,2018a; Yang et al., 2018). In the present study, bioinformatics analysiswas carried out based on the acquired unigenes data from five differenttissues of C. indicum. The genes of the main flavonoid pathways were

identified by KEGG, covering the majority of enzymes in key processessuch as flavonoid biosynthesis.

Based on data from our transcriptomic and enzyme activity assays, aputative linarin pathway was proposed. As illustrated in Fig. 7, enzymesinvolved in the proposed pathway encompass phenylalanine ammonia-lyase (PAL), trans-cinnamate 4-monooxygenase (C4H), 4-coumarate-CoA ligase (4CL), chalcone synthase (CHS), chalcone reductase (CHR),chalcone isomerase (CHI), FNSⅡ, flavonoid 3′-hydroxylase (F3′H),methyltransferase (MT), and UDP-glycosyltransferase (UGT). Eightcandidate unigenes of methyltransferases were found as putative onesinvolved in the catalysis of apigenin to acacetin. Acacetin and linarin(acacetin-β-rutinoside) have been detected in C. indicum and Chry-santhemum boreale (Gao et al., 2008; Nugroho et al., 2013). Further-more, acacetin-7-O-glucoside has been observed in Chrysanthemummorifolium cv. Hangju and Chrysanthemum morifolium cv. Kotobuki(Nishina et al., 2019; Zhang et al., 2019a). Therefore, UGTs might beinvolved in acacetin-7-O-glucoside biosynthesis, and this glucosidederivate might be further rhamnosylated by another UGT (or the sameUGT) to form linarin. In addition, apigenin-7-O-glucoside and luteolin-7-O-glucoside have been identified in C. indicum (Feng et al., 2010; Gaoet al., 2008; Wu et al., 2013). Therefore, UGTs, not yet identified, areprobably involved in flavone glycosides biosynthetic pathways. There-fore, the 53 flavonoid O-glucosyltransferase unigenes reported in ourtranscriptome have to be functionally characterized to demonstratetheir involvement in acacetin-7-O-glucoside, linarin, apigenin-7-O-glu-coside, or luteolin-7-O-glucoside biosynthesis.

Flavone synthase is a rate-limiting enzyme that catalyzes the bio-synthesis of flavones (Martens and Mithofer, 2005). The metabolic fluxfrom flavanones to the biosynthesis of flavones controlled by rice FNSⅡhas been demonstrated (Lam et al., 2014). Given the importance of FNSin flavonoid biosynthesis, RT-qPCR of a unigene 0043683 encodingFNSⅡ was performed with the highest transcript levels observed inflower and flower bud. This result is consistent with the expressionpattern of flavonoid transcriptome (Fig. 4). In other words, tran-scriptomic analysis has provided insights for elucidating the molecularsignatures of flavonoid accumulation in C. indicum.

Fig. 4. (a,b,c) RT-qPCR analysis of CiFNSⅡ ex-pression in different tissues. The relative mRNAexpression levels were calculated using the2−ΔΔCT method and were normalized toGAPDH, EF1α and the double references ofGAPDH and EF1α. Each tissue type was per-formed in three biological replicates. Error barsrepresent the standard deviations betweenbiological replicates. (d) RPKM values ofCiFNSⅡ expression in different tissues from thetranscriptome data.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

250

Two FNS enzymes, FNSⅠ ans FNSⅡ, have been shown to be re-sponsible for the conversion of flavanones to flavones by catalyzing adouble bond formation between C2 and C3 of flavanones. At present,most identified FNSⅡ enzymes are classified as CYP450 s, and arewidely distributed in dicot plants. On the other hand, FNSⅠ has beenreported in Arabidopsis thaliana, rice (Oryza sativa) and several speciesof Apiaceae (Ferreyra et al., 2015; Jiang et al., 2016; Wang et al.,2018b; Wu et al., 2016; Zhao et al., 2016). In our study, a C. indicumFNS with a higher expression in flower and flower bud was clusteredwith FNSⅡs from the other plants (Fig. 3b). C. indicum FNS is closelyrelated with FNSⅡs and distantly related with F2Hs. Thus, we predictedthat FNS in C. indicum is a FNS Ⅱ, and named it as CiFNSⅡ.

For enzyme assays, CiFNSⅡ was expressed in WAT11 yeast cells, andthe recombinant protein was demonstrated to convert flavanones(naringenin, eriodictyol, and liquiritigenin) into the correspondingflavones (apigenin, luteolin, and 7,4′-dihydroxyflavone) in vitro

(Fig. 5). When low amounts of substrates and total proteins were usedfor activity assays (10 μg and 50 μg, respectively), negative results wereachieved (Fig. S13). The CYP93B1 (F2H) enzyme from G. echinata hasbeen reported to catalyze the NADPH-dependent hydroxylation ofnaringenin and liquiritigenin to the corresponding 2eOH flavanones.Furthermore, eriodictyol is converted into luteolin by microsome ofyeast expressing CYP93B1 only after dehydration by acid treatment(Akashi et al., 1998). Currently, it have been reported that CYP450AFNS2 and TFNS5 from snapdragon and torenia, respectively, are FNSIIenzymes whose reaction mechanisms for flavone biosynthesis involve2-hydroxyflavanone intermediates (Akashi et al., 1999). The secondtype of reaction mechanism of FNSⅡ introducing a C2=C3 doublebond without 2-hydroxyflavanones formation has been observed inGerbera hydrida, G. max, L. japonica, and L. macranthoides (Fliegmannet al., 2010; Martens and Forkmann, 1999; Wu et al., 2016). In ourassays, 2-hydroxynaringenin (Fig. 5a) and 2-hydroxyliquiritigenin

Fig. 5. Analysis of reaction products generated by CiFNSⅡ recombinant protein by HPLC-MS. Chromatographic profiles show the conversion of (a) naringenin (NAR)to apigenin (API), (b) eriodictyol (ERI) to luteolin (LUT), and (c) liquiritigenin (LIQ) to and 7,4′-dihydroxyflavone (DHF). * represents a contamination peak.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

251

Fig. 6. Flavonoid contents (a) and HPLC-MS chromatographic profiles of (b) linarin and (c) luteolin in different tissues of C. indicum.

Table 2Flavonoids in different tissues of C. indicum.

Name Formula Mass [M−H]− RT Contents of flavonoids (μg/mg FW)

Root Stem Leaf Flower Flower bud

Luteolin C15H10O6 285.0405 9.57 0.069 ± 0.005 9.535 ± 0.072 4.230 ± 0.087 27.635 ± 0.055 18.324 ± 0.093Eriodictyol C15H12O6 287.0561 9.14 0.007 ± 0.001 0.002 ± 0.001 2.274 ± 0.03 0.294 ± 0.014 0.57 ± 0.002Apigenin C15H10O5 269.0455 14.09 0.010 ± 0.002 0.244 ± 0.008 0.301 ± 0.012 2.353 ± 0.020 1.307 ± 0.025Naringenin C15H12O5 271.0612 13.75 n.d. n.d. 0.070 ± 0.005 0.179 ± 0.010 0.146 ± 0.0097,4′-Dihydroxy flavone C15H10O4 253.0506 8.92 n.d. n.d. n.d. n.d. n.d.Liquiritigenin C15H12O4 255.0663 7.36 n.d. n.d. n.d. n.d. n.d.Acacetin C16H12O5 283.0612 17.08 0.163 ± 0.012 n.d. n.d. 0.019 ± 0.001 n.d.Linarin C28H32O14 591.1719 8.33 0.364 ± 0.014 0.133 ± 0.012 0.070 ± 0.005 0.052 ± 0.001 0.064 ± 0.007

n.d., not detected.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

252

(Fig. 5c) could also be identified through the extracted ion peak. Ourresults suggest that CiFNSⅡ catalyzes the conversion of flavanones toflavones, probably through the generation of 2-hydroxyflavanone in-termediates. Additional experiments are needed to investigate the

catalytic mechanism of CiFNSⅡ.In our enzyme reaction assay, the extracted ion peak (271) in the

second plot (from top to bottom) of the panel (Fig. 5a) exhibited twopeaks. The reason why two peaks were present when the m/z search

Fig. 7. A schematic outline of flavonoidpathway proposed for C. indicum.PAL, phenylalanine ammonia-lyase;C4H, trans-cinnamate 4-mono-oxygenase; 4CL, 4-coumarate-CoA li-gase; CHS, chalcone synthase; CHR,chalcone reductase; CHI, chalcone iso-merase; F3′H, flavonoid 3′-hydroxylase;FNSⅡ, flavone synthase Ⅱ; MT, methyl-transferase; UGT, UDP-glycosyl-transferase. Enzymes highlighted in redrepresents the FNSⅡ that has been bio-chemically characterized, while com-pounds highlighted in green representsthe compounds detected in our study.The purple colored range shows theproposed pathway of flavanone to fla-vone identified in C. indicum accordingto the biochemical and phytochemicalresults. The solid arrows represent thecommon upstream pathway of flavo-noid biosynthesis in many plants(Artigot et al., 2013; Casas et al., 2014;Koes et al., 1994), the dashed arrowsrepresent proposed conversion in C. in-dicum (Nugroho et al., 2013).

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

253

corresponds to the specific product could be the O18 isotopic peakleading to the double peaks. The ratio of the peak areas in 271 to 269 is0.2483% (the red arrow in Fig. S14a,b). The ratio of the intensity of theions in 271 to 269 (Fig. S14c) is 0.2476%. The two values are similar tothe natural abundance of O18 (0.20%). Consequently, we concludedthat the confused peak of naringenin in Fig. 5b is the isotopic peak.

Traditionally, the dried flowers of C. indicum have been used for theextraction of metabolites to perform phytochemical and pharmacolo-gical analysis (Cha et al., 2018; Cheon et al., 2009; Matsuda et al., 2002;Seo et al., 2013; Zhang et al., 2019b). Different drying methods canaffect the chemical composition and abundance of metabolites (Abdul-Hamid et al., 2015; Farag et al., 2017; Pariyani et al., 2017). In ourstudy, the fresh root, stem, leaf, flower, and flower bud of C. indicumwere used to determine the levels of flavonones, flavones and theirderivatives in these different tissues and elucidate the metabolicpathway with the help of transcriptome. As a result, two flavonones(narigenin and eriodictyol), two flavones (apigenin and luteolin) andtwo flavone derivatives (acacetin and linarin) were identified in C. in-dicum. Unfortunately, liquiritigenin and 7,4′-dihydroxyflavone were notdetected in any tissues of C. indicum in our study. Recently, it has beenidentified in C. indicum flavonols and their derivatives such as quer-cetin, isorhamnetin, quercimeritroside and astroside (Cha et al., 2018;Luyen et al., 2015b). Future studies will be performed to compare thelevels of flavone and their derivatives with other accumulated flavo-noids, such as flavonols, anthocyanins and isoflavonoids in order tolearn more about flavonoids accumulated in C. indicum.

In Pharmacopoeia of the People’s Republic of China, linarin level isused as the standard for determining whether certain C. indicum pro-duct is qualified for medicinal use (Zhang et al., 2007). Notably, aprecursor of linarin (acacetin-7-O-rutinoside), acacetin has been re-ported to have anti-inflammatory, antioxidative and cardioprotectiveactivities (Kim et al., 2016; Liu et al., 2016; Sun et al., 2017). In ourstudy, both acacetin and linarin were highly accumulated in root, withconcentrations of over 6-times higher than those in flower (Fig. 6b,Figs. S12d, S15). It is interesting that linarin levels are lower in leaf,flower, and flower bud than those in root and stem. Our hypothesis isthat C. indicum is a perennial plant, which means it dies during winterand grows back from its rootstock. If linarin synthesis occurs all yearround, this may explain higher accumulation levels of linarin (acacetin-β-rutinoside) observed in root and stem than flower, flower bud, andleaf. One possible explanation is that acacetin and linarin may useapigenin as a substrate or precursor, resulting in low accumulation ofapigenin in root and stem. Likelywise, luteolin levels are higher inflower and flower bud than those of the other tissues. It has also beenreported higher levels of flavonoid aglycones (luteolin and apigenin)than glycosides (linarin glycosylated acacetin) in Chrysanthemum mor-ifolium with yellow petals and Chrysanthemum morifolium cv. NobleWine with purple striped white petals (Han et al., 2017). The reasonwhy flavone aglycones are higher than linarin in C. indicum is worthfurther exploring. Luteolin, one of the most common flavones, exists invegetables and fruits such as celery, parsley, and peppermint. It hasbeen reported to exhibit antitumor, anti-inflammatory, and anti-obesityeffects (Kwon et al., 2018; Negro et al., 2012). These observations areconsistent with luteolin-rich artichoke, an important vegetable in theMediterranean (e.g., Italy, Spain, France, and North Africa) andAmerica diet (Kwon et al., 2018; Negro et al., 2012).

Recently, quercitrin present in C. indicum flowers has been reportedto reduce high-fat diet-induced obesity in mice (Cha et al., 2018; Nepaliet al., 2018). Quercitrin is a flavonol glycoside consisting in the flavonolquercetin and the deoxy sugar rhamnose. Quercetin is formed fromeriodictyol by flavanone 3-hydroxylase (F3H) and flavonol synthase(FLS) enzymes (Winkel-Shirley, 2001). In our transcriptome analysis,genes of five families of enzymes, namely F3H, FLS, dihydroflavonol 4-reductase (DFR), and glucosyltransferase (GT) involved in flavonol andanthocyanidin biosynthesis were also identified, respectively (TableS6). These unigenes displayed a cluster pattern (Fig. 2b) similar to the

common intermediates of flavonoid pathway, such as naringenin, api-genin, luteolin, and eriodictyol in the different tissues of C. indicum(Fig. 6a). If gene expression and metabolite accumulation are con-sistent, it could be hypothesized that F3H, FLS activities and quercetrinmight have similar accumulation patterns in C. indicum. Further studiesto explore this hypothesis are warranted.

In our phytochemical analysis, liquiritigenin and 7,4′-dihydroxy-flavone were not detected in any tissues of C. indicum. To date, liquir-itigenin and 7,4′-dihydroxyflavone have been found only in a fewplants, such as licorice and soybean (Akashi et al., 1998; Fliegmannet al., 2010). Both F2H from licorice and soybean are able to catalyzethe conversion of liquiritigenin to 7,4′-dihydroxyflavone (Akashi et al.,1998; Fliegmann et al., 2010): Although no unigene for CHR has beenfound in the transcriptome analysis of C. indicum, here, we demon-strated the enzymatic conversion of liquiritigenin to 7,4′-dihydroxy-flavone by CiFNSⅡ. Similarly, FNSⅡ proteins from L. japonica and L.macranthoides have been reported to catalyze the conversion of liquir-itigenin to 7,4′-dihydroxyflavone, although neither liquiritigenin nor7,4′-dihydroxyflavone were detected in these plants (Wu et al., 2016).In the root of Scutellaria baicalensis, FNSⅡ-2 uses pinocembrin ratherthan naringenin as an intermediate to produce chrysin (Zhao et al.,2016). In our studies, both naringenin and eriodictyol were detected inC. indicum by phytochemical analysis, and their conversions into api-genin and luteolin, respectively, by in vitro activity assays were con-firmed. As a result, our results suggest that CiFNSⅡ is an active flavonesynthase in planta involved in the conversion of naringenin and erio-dictyol to apigenin and luteolin, respectively, in planta. Future studieswith knock-out and knock-down plants are needed to verify this hy-pothesis. Finally, we propose a linarin biosynthesis pathway (Fig. 7)based on transcriptomic analysis, metabolic profiling, and enzyme ac-tivities of CiFNSⅡ. These results provide genomic resources for futureenhancement of linarin production using genetic strategies and meta-bolic engineering in C. indicum.

5. Conclusion

In this study, we report the use of transcriptomic analysis and in-tegrative metabolomics to reveal genes involved in flavone biosynthesisand flavonoid accumulation levels in C. indicum. It is clearly demon-strated that flavonoid distribution differs considerably among differenttissues of C. indicum. Based on transcriptomic data, CiFNSⅡ was iden-tified as the key enzyme responsible for flavone biosynthesis in C. in-dicum. We further confirmed the biochemical function of CiFNSⅡ inconverting naringenin, eriodictyol and liquiritigenin directly to theirrespective flavones. Our study provides insight into the potential ap-plication of molecular breeding and metabolic engineering for im-proving the quality of cultivated C. indicum.

Conflict of interest

The authors declare no conflict of interest.

Funding

We are grateful for the financial support of the start-up fund ofGuangzhou University of Chinese Medicine, Guangdong Pearl RiverTalents Plan (2017GC010368), key platform construction project ofDepartment of Education of Guangdong Province (Grant number2014KTSPT016) & Guangdong Science & Technology (Grant number2013B090600115).

Appendix A. Supplementary data

Supplementary material related to this article can be found, in theonline version, at doi:https://doi.org/10.1016/j.indcrop.2019.04.009.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

254

https://doi.org/10.1016/j.indcrop.2019.04.009

References

Abdul-Hamid, N.A., Abas, F., Ismail, I.S., Shaari, K., Lajis, N.H., 2015. Influence of dif-ferent drying treatments and extraction solvents on the metabolite profile and nitricoxide inhibitory activity of ajwa dates. J. Food Sci. 80, H2603–2611.

Akashi, T., Aoki, T., Ayabe, S., 1998. Identification of a cytochrome P450 cDNA encoding(2S)-flavanone 2-hydroxylase of licorice (Glycyrrhiza echinata L.; Fabaceae) whichrepresents licodione synthase and flavone synthase II. FEBS Lett. 431, 287–290.

Akashi, T., Fukuchi-Mizutani, M., Aoki, T., Ueyama, Y., Yonekura-Sakakibara, K., Tanaka,Y., Kusumi, T., Ayabe, S., 1999. Molecular cloning and biochemical characterizationof a novel cytochrome P450, flavone synthase Ⅱ, that catalyzes direct conversion offlavanones to flavones. Plant Cell Physiol. 40, 1182–1186.

Artigot, M.-P., Daydé, J., Berger, M., 2013. Expression of key genes of the isoflavonoidpathway in hypocotyls and cotyledons during soybean seed maturation. Crop Sci. 53,1096–1108.

Ayabe, S.-i., Akashi, T., 2006. Cytochrome P450s in flavonoid metabolism. Phytochem.Rev. 5, 271–282.

Britsch, L., Heller, W., Grisebach, H., 1981. Conversion of flavanone to flavone, dihy-droflavonol and flavonol with an enzyme system from cell cultures of parsley.Zeitschrift für Naturforschung C 36, 742–750.

Casas, M.I., Duarte, S., Doseff, A.I., Grotewold, E., 2014. Flavone-rich maize: an oppor-tunity to improve the nutritional value of an important commodity crop. Front. PlantSci. 5, 440.

Cha, J.Y., Nepali, S., Lee, H.Y., Hwang, S.W., Choi, S.Y., Yeon, J.M., Song, B.J., Kim, D.K.,Lee, Y.M., 2018. Chrysanthemum indicum L. ethanol extract reduces high-fat diet-in-duced obesity in mice. Exp. Ther. Med. 15, 5070–5076.

Chapple, C., 1998. Molecular-genetic analysis of plant cytochrome P450-dependentmonooxygenases. Annu. Rev. Plant Biol. 49, 311–343.

Cheon, M.S., Yoon, T., Lee, D.Y., Choi, G., Moon, B.C., Lee, A.Y., Choo, B.K., Kim, H.K.,2009. Chrysanthemum indicum Linne extract inhibits the inflammatory response bysuppressing NF-kappaB and MAPKs activation in lipopolysaccharide-induced RAW264.7 macrophages. J. Ethnopharmacol. 122, 473–477.

Dixon, R.A., Pasinetti, G.M., 2010. Flavonoids and isoflavonoids: from plant biology toagriculture and neuroscience. Plant Physiol. 154, 453–457.

Dong, M., Yu, D., Abdullah Al-Dhabi, N., Duraipandiyan, V., 2017. The impacts ofChrysanthemum indicum extract on oxidative stress and inflammatory responses inadjuvant-induced arthritic rats. Evid. Based Complement. Altern. Med. 2017.

Falcone Ferreyra, M.L., Rius, S.P., Casati, P., 2012. Flavonoids: biosynthesis, biologicalfunctions, and biotechnological applications. Front. Plant Sci. 3, 222.

Farag, M.A., Ali, S.E., Hodaya, R.H., El-Seedi, H.R., Sultani, H.N., Laub, A., Eissa, T.F.,Abou-Zaid, F.O.F., Wessjohann, L.A., 2017. Phytochemical profiles and antimicrobialactivities of Allium cepa Red cv. and A. sativum subjected to different drying methods:a comparative MS-based metabolomics. Molecules 22.

Feng, Z., Yang, Y., Jiang, J., Zhang, P., 2010. Chemical constituents from flowers ofChrysanthemum indicum. China J. Chinese Matera Med. 35, 3302–3305.

Ferreyra, M.L.F., Emiliani, J., Rodriguez, E.J., Campos-Bermudez, V.A., Grotewold, E.,Casati, P., 2015. The identification of maize and Arabidopsis type I flavone synthaseslinks flavones with hormones and biotic interactions. Plant Physiol. 169, 1090–1107.

Fliegmann, J., Furtwangler, K., Malterer, G., Cantarello, C., Schuler, G., Ebel, J., Mithofer,A., 2010. Flavone synthase Ⅱ (CYP93B16) from soybean (Glycine max L.).Phytochemistry 71, 508–514.

Gao, M.H., Li, H., Zhang, L., Xiao, S.X., 2008. Studies on chemical constituents fromflowers of Chrysanthemum indicum. J. Chin. Med. Mater. 31, 682–684.

Gharbi, S., Shamsara, M., Khateri, S., Soroush, M.R., Ghorbanmehr, N., Tavallaei, M.,Nourani, M.R., Mowla, S.J., 2015. Identification of reliable reference genes forquantification of microRNAs in serum samples of sulfur mustard-exposed veterans.Cell J. (Yakhteh) 17, 494.

Gietz, R.D., Woods, R.A., 2002. Transformation of yeast by lithium acetate/single-stranded carrier DNA/polyethylene glycol method. Methods Enzymol 350, 87–96.

Gu, C., Chen, S., Liu, Z., Shan, H., Luo, H., Guan, Z., Chen, F., 2011. Reference geneselection for quantitative real-time PCR in Chrysanthemum subjected to biotic andabiotic stress. Mol. Biotechnol. 49, 192.

Haas, B.J., Papanicolaou, A., Yassour, M., Grabherr, M., Blood, P.D., Bowden, J., Couger,M.B., Eccles, D., Li, B., Lieber, M., 2013. De novo transcript sequence reconstructionfrom RNA-seq using the Trinity platform for reference generation and analysis. Nat.Protoc. 8, 1494.

Han, A.R., Kim, H.Y., So, Y., Nam, B., Lee, I.S., Nam, J.W., Jo, Y.D., Kim, S.H., Kim, J.B.,Kang, S.Y., Jin, C.H., 2017. Quantification of antioxidant phenolic compounds in anew Chrysanthemum Cultivar by high-performance liquid chromatography with diodearray detection and electrospray ionization mass spectrometry. Int. J. Anal. Chem.2017, 1254721.

Hasemann, C.A., Kurumbail, R.G., Boddupalli, S.S., Peterson, J.A., Deisenhofer, J., 1995.Structure and function of cytochromes P450: a comparative analysis of three crystalstructures. Structure 3, 41–62.

Jeong, S.C., Kim, S.M., Jeong, Y.T., Song, C.H., 2013. Hepatoprotective effect of waterextract from Chrysanthemum indicum L. flower. Chin. Med. 8, 7.

Jiang, N., Doseff, A.I., Grotewold, E., 2016. Flavones: from biosynthesis to health benefits.Plants (Basel) 5.

Kanehisa, M., Goto, S., Sato, Y., Furumichi, M., Tanabe, M., 2012. KEGG for integrationand interpretation of large-scale molecular data sets. Nucleic Acids Res. 40,D109–114.

Kim, J.H., Cheon, Y.M., Kim, B.-G., Ahn, J.-H., 2008. Analysis of flavonoids and char-acterization of the OsFNS gene involved in flavone biosynthesis in Rice. J. Plant Biol.51, 97.

Kim, C., Kim, M.C., Kim, S.M., Nam, D., Choi, S.H., Kim, S.H., Ahn, K.S., Lee, E.H., Jung,

S.H., 2013. Chrysanthemum indicum L. extract induces apoptosis through suppressionof constitutive STAT3 activation in human prostate cancer DU145 cells. Phytother.Res. 27, 30–38.

Kim, T.W., Kim, Y.J., Park, S.R., Seo, C.S., Ha, H., Shin, H.K., Jung, J.Y., 2015.Chrysanthemum indicum attenuates cisplatin-induced nephrotoxicity both in vivo andin vitro. Nat. Prod. Commun. 10, 397–402.

Kim, B., Lee, J.H., Seo, M.J., Eom, S.H., Kim, W., 2016. Linarin down-regulates phago-cytosis, pro-inflammatory cytokine production, and activation marker expression inRAW264.7 macrophages. Food Sci. Biotechnol. 25, 1437–1442.

Kochs, G., Grisebach, H., 1987. Induction and characterization of a NADPH-dependentflavone synthase from cell cultures of soybean. Zeitschrift für Naturforschung C 42,343–348.

Koes, R.E., Quattrocchio, F., Mol, J.N., 1994. The flavonoid biosynthetic pathway inplants: function and evolution. BioEssays 16, 123–132.

Kumar, S., Stecher, G., Tamura, K., 2016. MEGA7: molecular evolutionary geneticsanalysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874.

Kwon, E.Y., Kim, S.Y., Choi, M.S., 2018. Luteolin-enriched artichoke leaf extract alleviatesthe metabolic syndrome in mice with high-fat diet-induced obesity. Nutrients 10.

Lam, P.Y., Zhu, F.Y., Chan, W.L., Liu, H., Lo, C., 2014. Cytochrome P450 93G1 is a flavonesynthase Ⅱ that channels flavanones to the biosynthesis of tricin O-Linked conjugatesin rice. Plant Physiol. 165, 1315–1327.

Lee, H.J., Seo, H.S., Ryu, J., Yoon, Y.P., Park, S.H., Lee, C.J., 2015. Luteolin inhibited thegene expression, production and secretion of MUC5AC mucin via regulation of nu-clear factor kappa B signaling pathway in human airway epithelial cells. Pulm.Pharmacol. Ther. 31, 117–122.

Liu, H., Yang, L., Wu, H.J., Chen, K.H., Lin, F., Li, G., Sun, H.Y., Xiao, G.S., Wang, Y., Li,G.R., 2016. Water-soluble acacetin prodrug confers significant cardioprotectionagainst ischemia/reperfusion injury. Sci. Rep. 6, 36435.

Liu, X., Cheng, J., Zhang, G., Ding, W., Duan, L., Yang, J., Kui, L., Cheng, X., Ruan, J., Fan,W., Chen, J., Long, G., Zhao, Y., Cai, J., Wang, W., Ma, Y., Dong, Y., Yang, S., Jiang,H., 2018. Engineering yeast for the production of breviscapine by genomic analysisand synthetic biology approaches. Nat. Commun. 9, 448.

Luyen, B.T., Tai, B.H., Thao, N.P., Cha, J.Y., Lee, H.Y., Lee, Y.M., Kim, Y.H., 2015a. Anti-inflammatory components of Chrysanthemum indicum flowers. Bioorg. Med. Chem.Lett. 25, 266–269.

Luyen, B.T., Tai, B.H., Thao, N.P., Lee, Y.M., Lee, S.H., Jang, H.D., Kim, Y.H., 2015b. Theanti-osteoporosis and antioxidant activities of chemical constituents fromChrysanthemum indicum flowers. Phytother. Res. 29, 540–548.

Mahmood, T., Yang, P.-C., 2012. Western blot: technique, theory, and trouble shooting.N. Am. J. Med. Sci. 4, 429.

Martens, S., Forkmann, G., 1999. Cloning and expression of flavone synthase Ⅱ fromGerbera hybrids. Plant J. 20, 611–618.

Martens, S., Mithofer, A., 2005. Flavones and flavone synthases. Phytochemistry 66,2399–2407.

Mathesius, U., 2018. Flavonoid Functions in Plants and Their Interactions with OtherOrganisms. Multidisciplinary Digital Publishing Institute.

Matsuda, H., Morikawa, T., Toguchida, I., Harima, S., Yoshikawa, M., 2002. Medicinalflowers. VI. Absolute stereostructures of two new flavanone glycosides and a phe-nylbutanoid glycoside from the flowers of Chrysanthemum indicum L.: their inhibitoryactivities for rat lens aldose reductase. Chem. Pharm. Bull. 50, 972–975.

Negro, D., Montesano, V., Grieco, S., Crupi, P., Sarli, G., De Lisi, A., Sonnante, G., 2012.Polyphenol compounds in artichoke plant tissues and varieties. J. Food Sci. 77,C244–252.

Nepali, S., Cha, J.Y., Ki, H.H., Lee, H.Y., Kim, Y.H., Kim, D.K., Song, B.J., Lee, Y.M., 2018.Chrysanthemum indicum inhibits adipogenesis and activates the AMPK pathway inhigh-fat-diet-induced obese mice. Am. J. Chin. Med. 46, 119–136.

Nishina, A., Sato, D., Yamamoto, J., Kobayashi-Hattori, K., Hirai, Y., Kimura, H., 2019.Antidiabetic-like effects of naringenin-7-O-glucoside from edible Chrysanthemum’Kotobuki’ and Naringenin by activation of the PI3K/Akt pathway and PPARγ. Chem.Biodivers. 16, e1800434.

Nugroho, A., Lim, S.C., Choi, J., Park, H.J., 2013. Identification and quantification of thesedative and anticonvulsant flavone glycoside from Chrysanthemum boreale. Arch.Pharm. Res. 36, 51–60.

Pariyani, R., Ismail, I.S., Ahmad Azam, A., Abas, F., Shaari, K., 2017. Identification of thecompositional changes in Orthosiphon stamineus leaves triggered by different dryingtechniques using 1H NMR metabolomics. J. Sci. Food Agric. 97, 4169–4179.

Schmittgen, T.D., Livak, K.J., 2008. Analyzing real-time PCR data by the comparative CTmethod. Nat. Protoc. 3, 1101–1108.

Seo, D.W., Cho, Y.R., Kim, W., Eom, S.H., 2013. Phytochemical linarin enriched in theflower of Chrysanthemum indicum inhibits proliferation of A549 human alveolar basalepithelial cells through suppression of the Akt-dependent signaling pathway. J. Med.Food 16, 1086–1094.

Shen, G.-M., Jiang, H.-B., Wang, X.-N., Wang, J.-J., 2010. Evaluation of endogenous re-ferences for gene expression profiling in different tissues of the oriental fruit flyBactrocera dorsalis (Diptera: Tephritidae). BMC Mol. Biol. 11, 76.

Stotz, G., Forkmann, G., 1981. Oxidation of flavanones to flavones with flower extracts ofAntirrhinum majus (snapdragon). Zeitschrift für Naturforschung C 36, 737–741.

Sun, H., Zhang, T., Fan, Q., Qi, X., Zhang, F., Fang, W., Jiang, J., Chen, F., Chen, S., 2015.Identification of floral scent in chrysanthemum cultivars and wild relatives by gaschromatography-mass spectrometry. Molecules 20, 5346–5359.

Sun, S., Jiang, P., Su, W., Xiang, Y., Li, J., Zeng, L., Yang, S., 2016. Wild chrysanthemumextract prevents UVB radiation-induced acute cell death and photoaging.Cytotechnology 68, 229–240.

Sun, L.C., Zhang, H.B., Gu, C.D., Guo, S.D., Li, G., Lian, R., Yao, Y., Zhang, G.Q., 2017.Protective effect of acacetin on sepsis-induced acute lung injury via its anti-in-flammatory and antioxidative activity. Arch. Pharm. Res. 1–12.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

255

http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0005http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0005http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0005http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0010http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0010http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0010http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0015http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0015http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0015http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0015http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0020http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0020http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0020http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0025http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0025http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0030http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0030http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0030http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0035http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0035http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0035http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0040http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0040http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0040http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0045http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0045http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0050http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0050http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0050http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0050http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0055http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0055http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0060http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0060http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0060http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0065http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0065http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0070http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0070http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0070http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0070http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0075http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0075http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0080http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0080http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0080http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0085http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0085http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0085http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0090http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0090http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0095http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0095http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0095http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0095http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0100http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0100http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0105http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0105http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0105http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0110http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0110http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0110http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0110http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0115http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0115http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0115http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0115http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0115http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0120http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0120http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0120http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0125http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0125http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0130http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0130http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0135http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0135http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0135http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0140http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0140http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0140http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0145http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0145http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0145http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0145http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0150http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0150http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0150http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0155http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0155http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0155http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0160http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0160http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0160http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0165http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0165http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0170http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0170http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0175http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0175http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0180http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0180http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0180http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0185http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0185http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0185http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0185http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0190http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0190http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0190http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0195http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0195http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0195http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0195http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0200http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0200http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0200http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0205http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0205http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0205http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0210http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0210http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0215http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0215http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0220http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0220http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0225http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0225http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0230http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0230http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0230http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0230http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0235http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0235http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0235http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0240http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0240http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0240http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0245http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0245http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0245http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0245http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0250http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0250http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0250http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0255http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0255http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0255http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0260http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0260http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0265http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0265http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0265http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0265http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0270http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0270http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0270http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0275http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0275http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0280http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0280http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0280http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0285http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0285http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0285http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0290http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0290http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0290

Urban, P., Mignotte, C., Kazmaier, M., Delorme, F., Pompon, D., 1997. Cloning, yeastexpression, and characterization of the coupling of two distantly related Arabidopsisthaliana NADPH-cytochrome P450 reductases with P450 CYP73A5. J. Biol. Chem.272, 19176–19186.

Wagner, G.P., Kin, K., Lynch, V.J., 2012. Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples. Theory Biosci. 131,281–285.

Wang, H., Ma, D., Yang, J., Deng, K., Li, M., Ji, X., Zhong, L., Zhao, H., 2018a. An in-tegrative volatile terpenoid profiling and transcriptomics analysis for gene miningand functional characterization of AvBPPS and AvPS involved in the monoterpenoidbiosynthesis in Amomum villosum. Front. Plant Sci. 9, 846.

Wang, Q.Z., Downie, S.R., Chen, Z.X., 2018b. Genome-wide searches and molecularanalyses highlight the unique evolutionary path of flavone synthase I (FNSI) inApiaceae. Genome 61, 103–109.

Winkel-Shirley, B., 2001. Flavonoid biosynthesis. A colorful model for genetics, bio-chemistry, cell biology, and biotechnology. Plant Physiol. 126, 485–493.

Wu, X.L., Li, C.W., Chen, H.M., Su, Z.Q., Zhao, X.N., Chen, J.N., Lai, X.P., Zhang, X.J., Su,Z.R., 2013. Anti-inflammatory effect of supercritical-carbon dioxide fluid extractfrom flowers and buds of Chrysanthemum indicum Linnen. Evid. Complement.Alternat. Med. 2013, 413237.

Wu, J., Wang, X.C., Liu, Y., Du, H., Shu, Q.Y., Su, S., Wang, L.J., Li, S.S., Wang, L.S., 2016.Flavone synthases from Lonicera japonica and L. macranthoides reveal differentialflavone accumulation. Sci. Rep. 6, 19245.

Xiao, J., Capanoglu, E., Jassbi, A.R., Miron, A., 2016. Advance on the flavonoid C-gly-cosides and health benefits. Crit. Rev. Food Sci. Nutr. 56, S29–S45.

Yan, J., Wang, B., Jiang, Y., Cheng, L., Wu, T., 2014. GmFNSⅡ-controlled soybean flavonemetabolism responds to abiotic stresses and regulates plant salt tolerance. Plant CellPhysiol. 55, 74–86.

Yang, H.M., Sun, C.Y., Liang, J.L., Xu, L.Q., Zhang, Z.B., Luo, D.D., Chen, H.B., Huang,Y.Z., Wang, Q., Lee, D.Y., Yuan, J., Li, Y.C., 2017a. Supercritical-carbon dioxide fluidextract from Chrysanthemum indicum enhances anti-tumor effect and reduces toxicityof bleomycin in tumor-bearing mice. Int. J. Mol. Sci. 18, 465.

Yang, W.S., Kim, D., Yi, Y.S., Kim, J.H., Jeong, H.Y., Hwang, K., Kim, J.H., Park, J., Cho,J.Y., 2017b. AKT-targeted anti-inflammatory activity of the methanol extract ofChrysanthemum indicum var. albescens. J. Ethnopharmacol. 201, 82–90.

Yang, Y., Jiang, Z., Guo, J., Yang, X., Xu, N., Chen, Z., Hao, J., Li, J., Pang, J., Shen, C.,2018. Transcriptomic analyses of Chrysanthemum morifolium Ramat under UV-B

radiation treatment reveal variations in the metabolisms associated with bioactivecomponents. Ind. Crops Prod. 124, 475–486.

Yoo, K.Y., Kim, I.H., Cho, J.H., Ahn, J.H., Park, J.H., Lee, J.C., Tae, H.J., Kim, D.W., Kim,J.D., Hong, S., Won, M.H., Kang, I.J., 2016. Neuroprotection of Chrysanthemum in-dicum Linne against cerebral ischemia/reperfusion injury by anti-inflammatory effectin gerbils. Neural Regen. Res. 11, 270–277.

Yoshikawa, M., Morikawa, T., Murakami, T., Toguchida, I., Harima, S., Matsuda, H.,1999. Medicinal flowers. I. Aldose reductase inhibitors and three new eudesmane-type sesquiterpenes, kikkanols A, B, and C, from the flowers of Chrysanthemum in-dicum L. Chem. Pharm. Bull. 47, 340–345.

Yoshikawa, M., Morikawa, T., Toguchida, I., Harima, S., Matsuda, H., 2000. Medicinalflowers. Ⅱ. Inhibitors of nitric oxide production and absolute stereostructures of fivenew germacrane-type sesquiterpenes, kikkanols D, D monoacetate, E, F, and Fmonoacetate from the flowers of Chrysanthemum indicum L. Chem. Pharm. Bull. 48,651–656.

Zhang, J., Subramanian, S., Zhang, Y., Yu, O., 2007. Flavone synthases from Medicagotruncatula are flavanone-2-hydroxylases and are important for nodulation. PlantPhysiol. 144, 741–751.

Zhang, X., Xie, Y.L., Yu, X.T., Su, Z.Q., Yuan, J., Li, Y.C., Su, Z.R., Zhan, J.Y., Lai, X.P.,2015. Protective effect of super-critical carbon dioxide fluid extract from flowers andbuds of Chrysanthemum indicum Linnen against ultraviolet-induced photo-aging inmice. Rejuvenation Res. 18, 437–448.

Zhang, N., He, Z., He, S., Jing, P., 2019a. Insights into the importance of dietary chry-santhemum flower (Chrysanthemum morifolium cv. Hangju)-wolfberry (Lycium bar-barum fruit) combination in antioxidant and anti-inflammatory properties. Food Res.Int. 116, 810–818.

Zhang, X., Wu, J.Z., Lin, Z.X., Yuan, Q.J., Li, Y.C., Liang, J.L., Zhan, J.Y., Xie, Y.L., Su,Z.R., Liu, Y.H., 2019b. Ameliorative effect of supercritical fluid extract ofChrysanthemum indicum Linnen against D-galactose induced brain and liver injury insenescent mice via suppression of oxidative stress, inflammation and apoptosis. J.Ethnopharmacol. 234, 44–56.

Zhao, Q., Zhang, Y., Wang, G., Hill, L., Weng, J.K., Chen, X.Y., Xue, H., Martin, C., 2016. Aspecialized flavone biosynthetic pathway has evolved in the medicinal plant,Scutellaria baicalensis. Sci. Adv. 2, e1501780.

Zheng, X., Luo, X., Ye, G., Chen, Y., Ji, X., Wen, L., Xu, Y., Xu, H., Zhan, R., Chen, W.,2015. Characterisation of two oxidosqualene cyclases responsible for triterpenoidbiosynthesis in Ilex asprella. Int. J. Mol. Sci. 16, 3564–3578.

Y. Jiang, et al. Industrial Crops & Products 134 (2019) 244–256

256

http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0295http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0295http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0295http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0295http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0300http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0300http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0300http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0305http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0305http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0305http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0305http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0310http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0310http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0310http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0315http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0315http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0320http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0320http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0320http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0320http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0325http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0325http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0325http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0330http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0330http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0335http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0335http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0335http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0340http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0340http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0340http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0340http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0345http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0345http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0345http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0350http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0350http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0350http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0350http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0355http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0355http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0355http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0355http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0360http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0360http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0360http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0360http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0365http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0365http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0365http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0365http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0365http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0370http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0370http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0370http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0375http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0375http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0375http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0375http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0380http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0380http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0380http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0380http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0385http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0385http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0385http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0385http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0385http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0390http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0390http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0390http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0395http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0395http://refhub.elsevier.com/S0926-6690(19)30259-6/sbref0395

Gene mining and identification of a flavone synthase II involved in flavones biosynthesis by transcriptomic analysis and targeted flavonoid profiling in Chrysanthemum indicum L.IntroductionMaterials and methodsMaterialsChemical sourcesPlant materials

RNA extraction, reverse transcription, and sequencingSequence de novo assembly and gene annotationDifferentially expressed gene analysis and screening of the unigenes involved in flavonoid biosynthesisCloning of CiFNSⅡ, sequence alignment and phylogenetic analysisRT-qPCR ExperimentsConstruction of expression vector and Western blot analysisIn vitro CiFNSⅡ activity assaysThe determination of flavonoid contents and HPLC-MS analysis

ResultsThe de novo assembled transcriptome of C. indicum and annotationIdentification of differentially expressed genesAnnotated genes involved in flavonoid biosynthesisIsolation of full-length ORF for CiFNSⅡ and sequence analysisExpression analysis of the CiFNSⅡ in different tissues of C. indicumCiFNSⅡ heterologous protein expression and in vitro enzyme activity assaysProfiling of Flavonoid Metabolites in different tissues of C. indicum

DiscussionConclusionConflict of interestFundingSupplementary dataReferences